[BostonCard]

Well, it is also possible (in fact almost certain) that the cut by the casino is large enough that it overcomes any systematic edge.  The odds aren't the true odds of a team winning, they are a payout ratio (in other words, if you bet $1 on Stanford you will get $41 if Stanford wins the National Championship).  By definition the sum of the odds of the field, converted into probabilities, should be equal to 100%, since one and only one team will win.  But if you convert the payout ratio to probability (if the odds are 40 to 1 as for Stanford, the implied probability is 1/(40 + 1) or 2.44%), and sum them up, any excess above 1 will be the Casino's take assuming that the betting comes in inversely proportional to the payout.  I went through the VegasInsider numbers, which has the odds for the top 50 or so teams.  Basically, the sum of the probabilities, excluding "the field" of the top 50 teams is 1.36.  That means that on average, the casino will make a 36+% return (even more, because the tail that wasn't included in my calculation).  Thus, the probability of any team winning is probably about 27% lower than the odds implies.  That gives the casino a fair bit of cushion in case there is some systematic error.  To give you an idea, for Ohio State odds are 15 to 4.  But it's normalized odds are more like 5 to 1.  Put another way, the casino set the odds to try to get 15.4% of the money to come in on Ohio State but it will make money if tOSU wins as long as less than 19.5% of the money is on Ohio State.

BC

[BostonCard]

Well, it is also possible (in fact almost certain) that the cut by the casino is large enough that it overcomes any systematic edge.  The odds aren't the true odds of a team winning, they are a payout ratio (in other words, if you bet $1 on Stanford you will get $41 if Stanford wins the National Championship).  By definition the sum of the odds of the field, converted into probabilities, should be equal to 100%, since one and only one team will win.  But if you convert the payout ratio to probability (if the odds are 40 to 1 as for Stanford, the implied probability is 1/(40 + 1) or 2.44%), and sum them up, any excess above 1 will be the Casino's take assuming that the betting comes in inversely proportional to the payout.  I went through the VegasInsider numbers, which has the odds for the top 50 or so teams.  Basically, the sum of the probabilities, excluding "the field" of the top 50 teams is 1.36.  That means that on average, the casino will make a 36+% return (even more, because the tail that wasn't included in my calculation).  Thus, the probability of any team winning is probably about 27% lower than the odds implies.  That gives the casino a fair bit of cushion in case there is some systematic error.  To give you an idea, for Ohio State odds are 15 to 4.  But it's normalized odds are more like 5 to 1.  Put another way, the casino set the odds to try to get 15.4% of the money to come in on Ohio State but it will make money if tOSU wins as long as less than 19.5% of the money is on Ohio State.

BC